IntroductionNonsense-mediated mRNA decay (NMD) is a quality-control
process found in all eukaryotic organisms studied to date
(Maquat, 2004a; Conti and Izaurralde, 2005). One role of this pro-
cess is to degrade mRNA harboring a premature termination
codon (PTC) to prevent the synthesis of truncated proteins that
could be nonfunctional or whose function may be deleterious to
cells. The NMD pathway has been shown to be involved in the
regulation of gene expression in yeast, Drosophila, and mam-
mals (Sureau et al., 2001; He et al., 2003; Mendell et al., 2004;
Wollerton et al., 2004; Rehwinkel et al., 2005).
In mammalian cells, NMD takes place after pre-mRNA
splicing and in most cases is mediated by a protein complex de-
posited 20–24 nucleotides upstream of exon–exon junctions
(Maquat, 2004b; Conti and Izaurralde, 2005; Lejeune and
Maquat, 2005). This protein complex called the exon junction
complex (EJC) is thought to recruit the evolutionarily conserved
UPF proteins that play an essential but still not fully character-
ized role in NMD. During what is referred to as the “pioneer
round of translation” (Ishigaki et al., 2001), PTCs are recognized
and the targeted mRNAs are degraded by both 5′–3′ decay
involving decapping and 5′–3′ exoribonucleases such as hXRN1
and hXRN2/hRAT1, and by 3′–5′ decay involving deadenyl-
ation and the exosome (Chen and Shyu, 2003; Lejeune et al.,
2003; Couttet and Grange, 2004).
NMD implicates the participation of hUPF proteins such
as hUPF1, hUPF2, hUPF3 (also named hUPF3a), and hUPF3X
(also called hUPF3b). The function of these hUPF proteins is
still unclear. However, it has been proposed that they are se-
quentially recruited by the EJC: hUPF3/3X fi rst, followed by
hUPF2, and fi nally hUPF1 in mammalian cells (Lykke-Andersen
et al., 2001; Kim et al., 2005). Interestingly, the function of
hUPF2 has been demonstrated to be dispensable in some NMD
cases, suggesting the existence of different pathways to elicit
NMD (Gehring et al., 2005).
UPF1 is a phosphoprotein that undergoes phosphoryl-
ation/dephosphorylation cycles during NMD (Page et al., 1999;
Pal et al., 2001; Ohnishi et al., 2003). UPF1 has been shown to
interact with release factors in yeast (Czaplinski et al., 1998) and
mammals (Kashima et al., 2006), and could link the EJC and
translation termination complex. A direct interaction between
hUPF1 and the cap-binding protein CBP80 has also recently
been demonstrated in mammalian cells (Hosoda et al., 2005),
indicating that hUPF1 establishes a complex interaction net-
work either before or during the pioneer round of translation.
Inhibition of nonsense-mediated mRNA decay (NMD) by a new chemical molecule reveals the dynamic of NMD factors in P-bodies
Sébastien Durand,1 Nicolas Cougot,1 Florence Mahuteau-Betzer,2 Chi-Hung Nguyen,2 David S. Grierson,2
Edouard Bertrand,1 Jamal Tazi,1 and Fabrice Lejeune1
1Centre National de la Recherche Scientifi que, Institut de Génétique Moléculaire de Montpellier, Université de Montpellier, Montpellier F-34293, France2Centre National de la Recherche Scientifi que, Laboratoire de Pharmacochimie, Institut Curie, Université Paris-Sud, Orsay F-91405, France
In mammals, nonsense-mediated mRNA decay (NMD)
is a quality-control mechanism that degrades mRNA
harboring a premature termination codon to prevent
the synthesis of truncated proteins. To gain insight into
the NMD mechanism, we identifi ed NMD inhibitor 1
(NMDI 1) as a small molecule inhibitor of the NMD path-
way. We characterized the mode of action of this com-
pound and demonstrated that it acts upstream of hUPF1.
NMDI 1 induced the loss of interactions between hSMG5
and hUPF1 and the stabilization of hyperphosphorylated
isoforms of hUPF1. Incubation of cells with NMDI 1 allowed
us to demonstrate that NMD factors and mRNAs subject
to NMD transit through processing bodies (P-bodies),
as is the case in yeast. The results suggest a model in
which mRNA and NMD factors are sequentially recruited
The online version of this article contains supplemental material.
JCB • VOLUME 178 • NUMBER 7 • 2007 1146
Phosphorylation of hUPF1 has been shown to be performed by
hSMG1, a phosphoinositide 3-kinase–related kinase (Page et al.,
1999; Pal et al., 2001; Yamashita et al., 2001), and requires the
presence of hUPF2 and hUPF3 (Kashima et al., 2006). In con-
trast, dephosphorylation of hUPF1 requires the presence of a
multiprotein complex composed of hSMG5, hSMG6, hSMG7,
and protein phosphatase 2A (Chiu et al., 2003; Ohnishi et al.,
2003). For the most part, hSMG5 and hSMG7 proteins are dis-
tributed evenly throughout the cytoplasm, but a fraction is also pre-
sent in processing bodies (P-bodies; Unterholzner and Izaurralde,
2004). hSMG6 is a cytoplasmic protein that concentrates in
cytoplasmic foci distinct from P-bodies and whose nature is
still unclear (Unterholzner and Izaurralde, 2004).
P-bodies have been described in lower and higher eukary-
otic cells (Ingelfi nger et al., 2002; van Dijk et al., 2002; Sheth
and Parker, 2003; Cougot et al., 2004). In mammals, these cyto-
plasmic structures contain many factors involved in mRNA de-
cay, including components of the decapping machinery such as
decapping protein 1a (DCP1a; Ingelfi nger et al., 2002), DCP2
(Ingelfi nger et al., 2002; van Dijk et al., 2002), GE1 (also called
Hedls; Fenger-Gron et al., 2005; Yu et al., 2005), p54/RCK
(Cougot et al., 2004), the deadenylase CCR4 (Cougot et al.,
2004), XRN1 (Bashkirov et al., 1997), the LSM1-7 complex
involved in different aspects of RNA processing (Cougot et al.,
2004; Ingelfi nger et al., 2002), and the hUPF1, hSMG5, and
hSMG7 components of the NMD machinery (Unterholzner and
Izaurralde, 2004; Fukuhara et al., 2005). The function of P-bodies
is still unclear but they may serve as a storage compartment for
both untranslated RNAs and proteins involved in RNA decay
(Brengues et al., 2005; Pillai et al., 2005; Teixeira et al., 2005;
Franks and Lykke-Andersen, 2007), and/or as a site for RNA
decay (Cougot et al., 2004; Sheth and Parker, 2006).
In a recent work, we showed that hydrophobic tetra-
cyclic indole derivatives block the function of specifi c splicing
factors (Soret et al., 2005). In light of these fi ndings, we de-
cided to look further at this collection to determine if certain of
these compounds also inhibit NMD. The underlying idea was
that such small molecule inhibitors could represent powerful
tools to decipher the NMD process. In this paper, we report
the identifi cation of one such molecule, nonsense-mediated
mRNA decay inhibitor 1 (NMDI 1), that inhibits nucleus-
associated as well as cytoplasmic NMD. The inhibitory mechanism
appears to be caused by the loss of the interaction between
hSMG5 and hUPF1, thereby leading to the stabilization of the
hyperphosphorylated forms of hUPF1 and to its concentra-
tion in P-bodies. Interestingly, NMDI 1–mediated inhibition
revealed that other NMD factors and PTC-containing mRNA
can traffi c through P-bodies as is the case in yeast (Sheth and
Parker, 2006).
ResultsIdentifi cation of a novel NMD inhibitorIn a previous paper we identifi ed a series of polycyclic indole
derivatives that block the function of specifi c splicing factors
(Soret et al., 2005). Because certain molecules from this family
inhibit the function of proteins involved in mRNA maturation,
we decided to assess their capacity to inhibit NMD. HeLa cells
were transfected by two test plasmids coding for β globin (Gl)
and glutathione peroxidase 1 (GPx1) mRNA, either with (Ter)
or without (Norm) a PTC. Gl mRNA was subject to nucleus-
associated NMD in nonerythroid cells (Thermann et al., 1998;
Zhang et al., 1998) whereas GPx1 mRNA was subject to cyto-
plasmic NMD (Moriarty et al., 1998). Additionally, a reference
plasmid coding for the mouse major urinary protein (MUP)
mRNA was also introduced into the cells (Ishigaki et al., 2001).
24 h after transfection, cells were incubated for 20 h with 5 μM
of indole compound (Table S1, available at http://www.jcb.org/
cgi/content/full/jcb.200611086/DC1) or DMSO(−) as a control.
Then, total RNAs were purifi ed and analyzed by RT-PCR
(Fig. 1) to measure NMD effi ciency. Among the 25 indole de-
rivatives tested, only compound 70 (NMDI 1) stabilized the
Gl Ter mRNA level about threefold, indicating that this mole-
cule is an inhibitor of nucleus-associated NMD (Fig. 1 A and
not depicted). Interestingly, NMDI 1 also stabilized the level of
GPx1 Ter mRNA by approximately twofold (Fig. 1 B and not
depicted). To confi rm these results, we measured the NMD in-
hibition by RNase protection assay (RPA), as it represents a
more reliable approach for RNA quantifi cation. The results are
presented in Fig. S1 (A and B) and confi rm the two- to threefold
NMD inhibition by NMDI 1. Altogether, these data allowed us to
conclude that NMDI 1 is an inhibitor of nucleus-associated as
well as cytoplasmic NMD. Notably, the inhibition level obtained
with NMDI 1 is similar to that observed with other NMD inhib-
itors such as cycloheximide (CHX) or to the down-regulation of
hDCP2 or hPARN (Ishigaki et al., 2001; Lejeune et al., 2003).
To show a more direct correlation between NMDI 1 and NMD
inhibition, we measured NMD effi ciency on PTC-containing
Gl or GPx1 mRNA in cells that were treated with an increasing
amount of NMDI 1 (Fig. 1 C). Interestingly, we observed a
progressive NMD inhibition from 0 to 5 μM NMDI 1 for both
Gl and GPx1 constructs. At >5 μM, we were unable to get a
substantially stronger inhibition, suggesting that NMD cannot be
100% eliminated in our experimental conditions or that the
20–30% of mRNA that escaped from NMD inhibition represents
the fraction of mRNAs already engaged in the NMD process at
the time of NMDI 1 treatment. In all subsequent experiments,
we used 5 μM NMDI 1 as our working concentration. Notably,
NMDI 1 does not exhibit any cellular toxicity, as measured by
trypan blue staining, even at concentrations as high as 125 μM
(unpublished data).
At this stage, some controls were performed to investigate
the specifi city of inhibition mediated by NMDI 1. First, NMDI 1
failed to have any effect on splicing of several pre-mRNA re-
porter transcripts (Soret et al., 2005) and did not affect the level
of pre-mRNA (Fig. S1, A and B). Second, unlike CHX, which
inhibits translation, NMDI 1 does not alter the expression of
transfected fi refl y luciferase (Fluc; Fig. 1 D), suggesting that
NMDI 1 is not a general translation inhibitor. To further dem-
onstrate the absence of any effects of NMDI 1 on translation,
we performed metabolic labeling of proteins with [35S]methionine
in HeLa cells and showed that treatment with NMDI 1 had no
effect on 35S incorporation (Fig. S1 C). Third, to assay the integ-
rity of the microRNA (miRNA) decay pathway in the presence
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1147
of NMDI 1, we used a Renilla luciferase (Rluc) construct that is
subject to degradation by let- 7 miRNA (pRL-Perf) or immune
to miRNA decay pathway (pRL-3XBugleMut; Pillai et al.,
2005). Our results indicate that NMDI 1 does not increase
Renilla activity, which is under the control of let-7 miRNA,
confi rming that targeted mRNA degradation by miRNA is not
altered by NMDI 1 (Fig. 1 E). Finally, we also tested whether
NMDI 1 could induce the formation of the stress granules that
provides a sensitive assay for proper mRNA metabolism.
Indeed, these structures are aggregates of messenger RNPs that
Figure 1. Identifi cation of an NMD inhibitor. (A) RT-PCR analysis of globin and MUP mRNA. 106 HeLa cells were transfected with a plasmid that codes for MUP mRNA and with test and reference plasmids that code for globin mRNA wild type (Norm) or harbor a PTC (Ter). After transfection, cells were incu-bated with DMSO(−) or 5 μM of a chemical compound for 20 h. Purifi ed RNA was reverse-transcribed to serve as a substrate for specifi c amplifi cation by PCR. The three leftmost lanes correspond to serial twofold dilutions of PCR template to ensure that the amplifi cation conditions are quantitative. (B) RT-PCR analysis of GPx1 mRNA wild type (Norm) or cells harboring a PTC (Ter). The experiment was performed as described in A. The measure of the level of Gl or GPx1 mRNA was normalized with the level of MUP mRNA. The level of each Gl or GPx1 Ter was normalized with the level of the corresponding Gl or GPx1 Norm and is reported as a percentage of Norm (number below each lane). (C) Dose-response effect of NMDI 1 on Gl or GPx1 Ter. Hela cells were transfected with pmCMV-Gl Ter or pmCMV-GPx1 Ter and with phCMV-MUP. After 24 h, cells were incubated with an increasing amount of NMDI 1. NMD was measured by quantitative radiolabeled RT-PCR and confi rmed by RPA. (D) Measure of Fluc activity. Cells were transfected with pRLuc and pFluc expression vectors, and then treated with DMSO(−), NMDI 1, or CHX. Fluc activity was measured by a luminometer and normalized according to the expression level of Fluc and Rluc mRNA. (E) NMDI 1 does not inhibit miRNA-induced mRNA decay. HeLa cells were transfected with either pRL-3XBulgeMut (RLm) or pRL-Perf (RLp; Pillai et al., 2005) and incubated for 24 h with DMSO(−) or NMDI 1. Histogram represents the ratio of Rluc/Fluc. Results were normalized to RLm, which was set at 10 arbitrary units. All results are representative of at least three independent experiments. Error bar denotes SD.
JCB • VOLUME 178 • NUMBER 7 • 2007 1148
form when cells are subjected to several stresses, including mild
translational inhibition. Unlike sodium arsenite treatment that is
commonly used to induce stress granule formation (Kedersha
et al., 2005), NMDI 1 treatment did not change the localization
of G3BP protein, a well-characterized marker of stress granules
(Fig S1 D; Tourriere et al., 2003). Collectively, these results in-
dicate that NMDI 1 is a new and specifi c NMD inhibitor.
NMDI 1 abrogates NMD upstream of hUPF1 functionsTo gain insight into the mode of inhibition of NMDI 1, we ana-
lyzed its effects on a tethering system that mimics the sequen-
tial recruitment of NMD factors on mRNA (Lykke-Andersen
et al., 2000; Kim et al., 2005). Cells were transfected with two
types of constructs. The fi rst codes for a Fluc mRNA contain-
ing eight binding sites for the MS2 protein in its 3′ untranslated
region and the second codes for either the MS2 protein or one
of the following fusions: MS2-hUPF1, MS2-hUPF2, or MS2-
hUPF3X. Additionally, we transfected HeLa cells with a con-
struct coding for the Rluc mRNA to normalize the amount of
analyzed RNA. Cells were then incubated for 20 h with NMDI 1
or DMSO(−) as a negative control, and Rluc as well as Fluc
mRNA levels were measured by RT-PCR as described previ-
ously (Hosoda et al., 2005). The expression of each MS2 fusion
was controlled by Western blot to verify that the observed ef-
fects were not caused by a variation in protein expression (Fig.
2 A). In each case, the compound did not affect expression of
the MS2 fusion, which was itself never higher than the level of
the endogenous protein. As expected, the control experiment
performed in the presence of DMSO revealed that the level
of Fluc mRNA was lower in cells expressing one of the MS2-
hUPF fusion proteins compared with cells expressing only MS2
(Fig. 2 B). Remarkably, NMDI 1 counteracted the degradation
induced by MS2-hUPF2 or MS2-hUPF3X but had no effect
against MS2-hUPF1 (Fig. 2 B). Notably, the inhibition levels
obtained with NMDI 1 were very similar to those observed
when NMD was inhibited through down-regulation of hCBP80
(Hosoda et al., 2005). To obtain a more accurate measure of the
NMD inhibition, Rluc and Fluc mRNA levels were also mea-
sured by RPA. The results are presented in Fig. S2 A (available
at http://www.jcb.org/cgi/content/full/jcb.200611086/DC1)
and reproduce the quantifi cation of mRNA levels by RT-PCR
(Fig. 2 B). Altogether, these results indicate that NMDI 1 inhibits
NMD downstream of hUPF3X or hUPF2 recruitment and up-
stream of hUPF1 functions.
NMDI 1 does not prevent the interactions between hUPF1 and hUPF3XIn the light of the results described in the previous paragraph,
we hypothesized that NMDI 1 could prevent the recruitment of
hUPF1 to the EJC via its interactions with the other hUPF
proteins. To test this, we immunoprecipitated hUPF1 from HeLa
cell extracts under conditions that preserve the integrity of mes-
senger RNPs (Lejeune and Maquat, 2004). NMDI 1 or DMSO(−)
was added to the cell culture 20 h before immunoprecipitation (IP).
Because hUPF2 was shown to be dispensable in some NMD
cases (Gehring et al., 2005), we focused our analysis on the
presence of hUPF3X protein in each IP (Fig. 3 A). As a control
for IP specifi city, we did not detect tubulin protein in any of the
hUPF1 IPs and no proteins were present in the IP performed
with normal rabbit serum. The results show that hUPF3X was
present in hUPF1 IP even when cells were incubated with our
NMD inhibitor. Thus, these data demonstrate that the interaction
between hUPF1 and hUPF3X is not abolished by NMDI 1 and
suggest that NMDI 1 would not prevent the recruitment of hUPF1
to the EJC.
NMDI 1 stabilizes hyperphosphorylated forms of hUPF1Because hUPF1 requires a cycle of phosphorylation and de-
phosphorylation during NMD (Ohnishi et al., 2003), we next
investigated the possibility that NMDI 1 may affect hUPF1
function by interfering with its phosphorylation level. We thus
measured the level of hUPF1 phosphorylation in cells incu-
bated with NMDI 1 or, as a control, with DMSO(−) by 2D gel
analysis. Because hUPF1 phosphorylation is influenced by
serum (Pal et al., 2001), we used 293T cells that, unlike HeLa
cells, can be synchronized by serum deprivation. Cells were
transfected with the expression vector pCI-neo-FLAG-hUpf1
Figure 2. NMDI 1 inhibits NMD before the functions of hUPF1. (A) HeLa cells were transiently transfected with plasmids that encode the Rluc mRNA, the Fluc mRNA containing MS2 binding sites in its 3′ untranslated region, and the mRNA coding for MS2 protein either alone or fused with one of the hUPF proteins. 24 h after transfection, HeLa cells were incubated either with DMSO(−) or NMDI 1 for 20 h. The expression level of each MS2 fu-sion protein was determined by Western blot. 10 μg of protein extract was loaded on 10% SDS–polyacrylamide gel and transferred to a nitrocellu-lose membrane before incubation with antibodies against each of the hUPF proteins. The position of endogenous and exogenous proteins is indicated on the right. (B) The level of Fluc mRNA was normalized with the level of the corresponding Rluc mRNA and is reported as a percentage of normal-ized Fluc mRNA when only MS2 protein was expressed. The three leftmost lanes correspond to serial twofold dilutions of PCR template to show that the amplifi cation conditions are quantitative.
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1149
(Sun et al., 1998) and synchronized for 24 h by serum depriva-
tion 12 h after transfection. Finally DMSO or 5 μM NMDI 1
was added for 3 h before adding back serum for 1 h. Our results
show that when serum was not added back to the cell culture,
the FLAG-hUPF1 protein electrofocalized in one spot corre-
sponding to the unphosphorylated protein (Fig. 3 B; Pal et al.,
2001). After serum addition, we observed a mild phosphoryl-
ation of FLAG-hUPF1 protein in the presence of DMSO and
the stabilization of hyperphosphorylated isoforms of FLAG-
hUPF1 when cells were incubated with NMDI 1 (Fig. 3 B).
We concluded that NMDI 1 stabilized hyperphosphorylated
isoforms of hUPF1.
As it has been proposed that hUPF1 would localize to
P-bodies when hyperphosphorylated (Unterholzner and Izaurralde,
2004; Fukuhara et al., 2005), we analyzed the cellular local-
ization of FLAG-hUPF1 in HeLa cells in the absence or pres-
ence of NMDI 1 (Fig. 3 C). With the exception of coexpression
experiments with hSMG7, which induces the recruitment of
hUPF1 to P-bodies (Fig. 3 C; Unterholzner and Izaurralde,
2004), exogenous hUPF1 was equally distributed through the
cytoplasm when cells were incubated with DMSO(−) as previ-
ously reported for untreated cells (Fig. 3 C; Lykke-Andersen
et al., 2000; Mendell et al., 2002). When cells were treated with
NMDI 1, we observed cytoplasmic concentrations of FLAG-
hUPF1 in some structures that colocalize with the three com-
monly used markers of P-bodies: GFP-GE1, YFP-hSMG7, or
CFP-hDCP1a (Fig. 3 C). We also verifi ed that FLAG-hUPF1
accumulated in P-bodies in the presence of NMDI 1 in 293T
cells under the same experimental conditions used to study the
phosphorylation level of FLAG-hUPF1 (Fig. S2 B). We used
hXRN1 protein as a P-body marker to avoid any additional
transfected DNA. After addition of serum, we observed some
FLAG-hUPF1 cytoplasmic concentrations that colocalize with
hXRN1 only when cells were treated with NMDI 1 but not in
its absence. To defi nitively demonstrate that hyperphosphory-
lated isoform of hUPF1 accumulates in P-bodies, HeLa cells
were treated for 20 h with either DMSO(−) or NMDI 1, and the
cellular localization of endogenous phosphorylated hUPF1
was determined using a specifi c antibody raised against a phos-
phoepitope of this protein (Ohnishi et al., 2003). The results
presented in Fig. S3 (available at http://www.jcb.org/cgi/content/
full/jcb.200611086/DC1) indicate that in the presence of NMDI 1,
phosphorylated hUPF1 isoforms colocalize with CFP-hDCP1a
foci. We conclude that NMDI 1 induces the accumulation of
hyperphosphorylated hUPF1 isoforms in P-bodies. This may
occur either via stimulation of phosphorylation or by blocking
dephosphorylation.
To distinguish between these two possibilities, we sub-
sequently investigated the association of hUPF1 with other NMD
partners in HeLa cells treated or untreated with NMDI 1 (Fig.
3 A). We fi rst analyzed the interaction of hUPF1 with its dephos-
phorylation complex. Immunoprecipitation of hUPF1 allowed
recovery of hSMG5, hSMG6, and hSMG7 from DMSO-treated
cells. However, after treatment of HeLa cells with NMDI 1
only hSMG6 and hSMG7 but not hSMG5 were still associated
with hUPF1 (Fig. 3 A). Thus, we conclude that NMDI 1 destabi-
lizes the interaction between hUPF1 and hSMG5. The fact that
NMDI 1 does not alter the association of hUPF1 with hSMG1
and hUPF3X strongly suggests that NMDI 1 does not infl uence
the interactions between hUPF1 and its phosphorylation com-
plex (Fig. 3 A). Altogether, our results indicate that the hyper-
phosphorylation of hUPF1 is most likely caused by a failure in
dephosphorylation because of the loss of interaction between
hUPF1 and hSMG5 rather than an activation of phosphor-
ylation. This conclusion is consistent with fi ndings showing
that hSMG5 is essential for hUPF1 dephosphorylation (Ohnishi
et al., 2003).
hSMG5 is excluded from P-bodies in the presence of NMDI 1Because NMDI 1 induces the localization of hUPF1 in P-bodies,
hUPF1 hyperphosphorylation, and the destabilization of inter-
actions between hUPF1 and hSMG5, we assessed the cellular
localization of the hUPF1 dephosphorylation complex during
NMDI 1 treatment. hSMG5 and hSMG7 have been shown to
localize mainly in the cytoplasm and particularly in P-bodies
as shown by colocalization experiments with the endogenous
LSM4 for hSMG7 and with hSMG7 for hSMG5 (Unterholzner
and Izaurralde, 2004). hSMG6 similarly localizes mainly in the
cytoplasm and also in some cytoplasmic foci that do not con-
tain endogenous LSM4 (Unterholzner and Izaurralde, 2004).
We transfected HeLa cells with expression vectors encoding
YFP-hSMG5, YFP-hSMG6, or YFP-hSMG7 (Unterholzner
and Izaurralde, 2004) and CFP-hDCP1a as a P-body marker.
After 24 h, we added DMSO(−) or 5 μM NMDI 1 to the cells.
As previously shown, in the absence of the inhibitor, YFP-
hSMG5, YFP-hSMG6, and YFP-hSMG7 were concentrated in
cytoplasmic foci (Unterholzner and Izaurralde, 2004; Fukuhara
et al., 2005), which, for a substantial fraction of them (33, 72,
and 100%, respectively), colocalized with CFP-hDCP1a (Fig.
3 D). The fact that we observed hSMG6 in P-bodies unlike what
was previously observed (Unterholzner and Izaurralde, 2004)
was likely caused by the different markers used for detection of
P-bodies and may refl ect heterogeneity of P-bodies in their pro-
tein composition (see Discussion). In the presence of NMDI 1,
the cytoplasmic foci containing YFP-hSMG6 or YFP-hSMG7
perfectly colocalized with CFP-hDCP1a P-bodies. Interest-
ingly, hSMG5 was no longer observed in cytoplasmic foci and
became evenly distributed in the cytoplasm in cells treated with
NMDI 1 (Fig. 3 D).
Endogenous hSMG5, hSMG6, or hSMG7 cannot be de-
tected in cytoplasmic foci because of their weak expression
(Unterholzner and Izaurralde, 2004). Because NMDI 1 inhibits
NMD and induces the accumulation of hUPF1 in P-bodies as
shown with exogenous as well as endogenous hUPF1 (Figs. 3 C,
S2 B, and S3), we tested whether the three hSMG proteins
would also accumulate in cytoplasmic foci of treated cells.
The results shown in Fig. S3 indicated that these three pro-
teins were not detected in cytoplasmic foci in DMSO-treated
cells. However, when cells were incubated with NMDI 1,
both hSMG6 and hSMG7 colocalized with CFP-hDCP1a in
P-bodies. In agreement with the transfection experiment (Fig.
3 D) under the same conditions, hSMG5 was not detected in
cytoplasmic foci. Altogether, our results indicate that NMDI 1
JCB • VOLUME 178 • NUMBER 7 • 2007 1150
Figure 3. NMDI 1 modifi es the cellular distribution of hUPF proteins and stabilizes hyperphosphorylated isoforms of FLAG-hUPF1. (A) Endogenous hUPF1 protein was immunopurifi ed using rabbit anti-hUPF1 antibodies from HeLa cell extracts that were incubated with DMSO(−) or 5 μM NMDI 1(+). In parallel, an immunopurifi cation control was performed using normal rabbit serum to verify the specifi city of the immunopurifi cation protocol. The three leftmost lanes cor-respond to serial twofold dilutions of a whole HeLa cell extract. The asterisk marks an uncharacterized band that presumably represents a degradation product. (B) 2D gel analysis of the FLAG-hUPF1 phosphorylation level. 106 293T cells were transfected with 1 μg pCI-neo-FLAG-hUPF1 plasmid. 12 h later, the serum was removed from the culture medium for 24 h, and then either DMSO or 5 μM NMDI 1 was added to the culture medium for 3 h before adding back 10% serum (sample (−+) except in sample (−−)). Proteins were extracted and loaded on a 2D gel analysis system (see Materials and methods). (C) Immunofl uorescence assay. HeLa cells were transfected with pCI-neo-FLAG-hUpf1 and either pYFP-hSmg7, pGFP-Ge1, or pCFP-hDcp1a. 24 h after transfection, cells were incubated with DMSO(−) or 5 μM NMDI 1 for 20 h. The blue staining visible in the merge of the two leftmost set of images corresponds to nuclei staining by Hoechst stain. (D) Immunofl uorescence assay. HeLa cells were transfected with pYFP-hSmg5, pYFP-hSmg6, or pYFP-hSmg7 and with pCFP-hDcp1a expression vectors. 24 h after transfection, cells were cultured in the presence of DMSO(−) or 5 μM NMDI 1 for 20 h. The white squares are magnifi cations of cell areas. Bars, 10 μm.
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1151
modifi es the cellular localization of hSMG5 by excluding it
from P-bodies. This is consistent with the failure of hUPF1 anti-
bodies to immunoprecipitate hSMG5 from NMDI 1–treated
cells (Fig. 3 A).
hUPF3 and hUPF3X localize in P-bodies when NMD is blocked by NMDI 1Because some NMD factors such as hUPF1, hSMG5, hSMG6,
or hSMG7 localize to P-bodies (Unterholzner and Izaurralde,
2004, this paper) we envisaged that other NMD factors may
pass through P-bodies in a transient manner. As NMDI 1 is able
to block NMD at a step where hUPF1 is confi ned to P-bodies,
we investigated the cellular localization of hUPF3 and hUPF3X
in both treated and untreated cells. These two proteins have been
previously shown to be primarily nuclear proteins in untreated
cells (Serin et al., 2001). We transfected HeLa cells with
expression vectors that code for hUPF3-FLAG or hUPF3X-
FLAG together with one of the P-body markers YFP-hSMG6,
YFP-hSMG7, GFP-GE1, or CFP-hDCP1a. The cells were
treated with DMSO or NMDI 1 before we performed indirect
immunofl uorescence experiments (Fig. 4 and not depicted for
CFP-hDCP1a). As for untreated cells (Lykke-Andersen et al.,
2000; Serin et al., 2001), hUPF3 or hUPF3X localized primar-
ily in the nucleus when cells were incubated with DMSO(−)
(Fig. 4). However, when cells were grown in the presence
of NMDI 1, we observed a cytoplasmic localization of hUPF3
as well as hUPF3X, with some accumulation in foci. To fur-
ther characterize these foci, we analyzed their colocalization
with cotransfected P-body markers (Fig. 4 and not depicted).
hUPF3-FLAG as well as hUPF3X-FLAG proteins colocalized
with YFP-hSMG6 (68 and 57%, respectively) or YFP-hSMG7
(50 and 51%, respectively) in cells treated with NMDI 1.
Surprisingly, unlike with FLAG-hUPF1, GFP-GE1 did not co-
localize with hUPF3-FLAG or hUPF3X-FLAG. We conclude
that hUPF3/3X proteins can translocate to the cytoplasm,
which is consistent with their shuttling properties (Lykke-
Andersen et al., 2000; Serin et al., 2001), and can reach a
subset of P-bodies.
Figure 4. hUPF3 localizes to P-bodies when NMD is inhibited. HeLa cells were transfected with a pcDNA3-hUpf3-FLAG or pcDNA3-hUpf3X-FLAG plasmid and either pGFP-Ge1, pYFP-hSmg6, or pYFP-hSmg7. 24 h after transfection, cells were incubated either with DMSO(−) or with 5 μM NMDI 1 for 20 h. In the merge, nuclei are visualized by Hoechst staining in blue. The white squares are magnifi cations of cell areas. Bars, 10 μm.
JCB • VOLUME 178 • NUMBER 7 • 2007 1152
PTC-containing mRNAs accumulate in P-bodies in the presence of NMDI 1Because NMD factors accumulate in P-bodies when NMD is
inhibited, we were interested in determining whether NMD sub-
strates also accumulate in P-bodies. In yeast, it has recently been
shown that PTC-containing mRNAs accumulate in P-bodies
when NMD is blocked (Sheth and Parker, 2006). We speculated
that our NMD inhibitor would allow us to reach the same con-
clusion in mammalian cells. HeLa cells were transfected with
pmCMV-Gl Ter or pmCMV-GPx1 Ter and the localization
of the resulting mRNAs was analyzed with several P-body
(Figs. 5 and S4, available at http://www.jcb.org/cgi/content/full/
jcb.200611086/DC1). As expected, in the absence of the inhibitor
we were unable to detect PTC-containing mRNAs, most likely
because of their rapid decay by NMD. However, in the presence
of NMDI 1, PTC-containing mRNAs were stabilized and de-
tected mainly in cytoplasmic aggregates. Interestingly, our results
indicated a substantial colocalization between PTC-containing
Gl or GPx1 mRNA and each tagged version of the tested hUPF
proteins CFP-hDCP1a, YFP-hSMG6, or YFP-hSMG7 (Figs. 5 A
and S4). Although the colocalization between Gl Ter or GPx1
Ter mRNAs and hUPF3/3X was total (Fig. 5 and not depicted for
hUPF3X), it was only partial with other P-body components such
as FLAG-hUPF1 (63 and 74%, respectively), YFP-hSMG6 (76
and 81%, respectively), or YFP-hSMG7 (71 and 94%, respec-
tively), and infrequently with GFP-hCCR4 (11 and 33%, respec-
tively; Figs. 5 A and S4). Notably, we often observed that RNA
foci and P-bodies did not overlap perfectly. Additionally, we
were unable to observe a colocalization between PTC-containing
mRNAs and GFP-GE1 (Figs. 5 A and S4). Altogether, our results
show that PTC-containing mRNAs were present in and adjacent
to P-bodies when NMD was inhibited by NMDI 1. In addition,
these data indicate heterogeneity in the composition of P-bodies
because some markers colocalize with PTC-containing mRNAs
while others do not.
The accumulation of PTC-containing mRNAs in P-bodies
when NMD is blocked in mammalian cells was confi rmed by
a more resolutive approach in U2OS cells. In this setting, we
tagged the mRNA with a 24× MS2 binding site repeat because
this approach can effi ciently detect single mRNA molecules by in
situ hybridization (Fusco et al., 2003; Fig. 5 B). In control cells,
PTC-containing mRNAs were mostly detected in the nucleus,
and the cytoplasmic molecules that were detected did not accu-
mulate in P-bodies labeled with CFP-hDCP1a. When NMD was
inhibited with NMDI 1, higher levels of Gl-Ter MS2 mRNAs
were detected in the cytoplasm, and these molecules accumulated
in structures that colocalized with P-bodies. As previously ob-
served, mRNAs did not perfectly colocalize with CFP-hDCP1a,
but were instead adjacent and formed a ring at the periphery of
P-bodies, similar to what was found with miRNA targets (Pillai
et al., 2005). These data confi rmed that mRNAs subjected to
NMD accumulate in P-bodies when their degradation is inhibited,
and this conclusion seems not be cell type specifi c.
We also analyzed the localization of Gl or GPx1 Norm
mRNAs in HeLa and U2OS cells treated with NMDI 1 (Fig. 6).
Our results showed no specifi c accumulation of these mRNAs
in cytoplasmic bodies typifi ed by YFP-hSMG6, GFP-GE1, or
CFP-hDCP1a in the presence of NMDI 1. In contrast, we did
not see any wild-type mRNAs in P-bodies either because Norm
mRNAs do not go to P-bodies or because their degradation
pathway is not affected by either DMSO or NMDI 1. These
results support the idea that NMDI 1 is an NMD inhibitor rather
than a general RNA decay inhibitor.
Only some steps of the NMD process occur in P-bodiesTo determine whether the accumulation of NMD factors in
P-bodies can be triggered when any step of the NMD process
is inhibited, we decided to interfere with the NMD process in
three different ways. The fi rst one relies on the down-regulation
of hUPF2 using siRNA (Kim et al., 2005; Fig. 7 A). According to
the current model of NMD in mammalian cells (Maquat, 2004b),
depletion of hUPF2 will block NMD at an earlier step than the
one induced by NMDI 1. Interestingly, hUPF2 down-regulation
does not induce the accumulation of FLAG-hUPF1, hUPF3-
FLAG, or hUPF3X-FLAG P-bodies. Additionally, whereas
hUPF2 depletion induces a stabilization of Gl-Ter mRNA be-
cause of the inhibition of NMD, Gl-Ter mRNA was homo-
genously distributed in the cytoplasm with no accumulation in
P-bodies (Fig. 7 B). Thus, inhibition of NMD by hUPF2 deple-
tion does not trigger accumulation of NMD factors and sub-
strates in P-bodies.
The second one is based on the down-regulation by siRNA
of a protein involved in a late step of the NMD process such as
hXRN1 (Cougot et al., 2004; Fig. 7 C). As for the down-regulation
of hUPF2, the cellular localization of NMD factors including
FLAG-hUPF1 or hUPF3-FLAG was not modifi ed by the cellu-
lar lack of the hXRN1 protein (Fig. 7 D), whereas Gl-Ter RNA
was detected in P-bodies. This result may indicate that the re-
cycling of NMD factors had already occurred before the function
of hXRN1. We conclude that the presence of NMD factors in
P-bodies depends on the NMD step, i.e., only some steps in the
NMD process occur in P-bodies.
According to our results, hUPF1 dephosphorylation is
one of the NMD steps that occur in P-bodies. In the presence
of NMDI 1, NMD is inhibited because hUPF1 is stalled in a
hyperphosphorylated form caused by the release of hSMG5
from P-bodies. To further confi rm this model, we aimed to
mimic the effect of NMDI 1 by analyzing the cellular local-
ization of NMD factors in the presence of the hUPF1 mutant
protein (HA-hUPF1dNT) that has been shown to prevent its
interaction with hSMG5 because it lacks an N- terminal, or in
the presence of hSMG5 mutant proteins (HA-hSMG5dCT
and HA-hSMG5DA) that cannot dephosphorylate hUPF1 be-
cause of either the lack of a C-terminal or the substitution
of aspartic acid 860 by an alanine (Ohnishi et al., 2003). As
shown in Fig. 8 A, when HeLa cells express HA-hUPF1dNT
protein, hUPF3-FLAG, hUPF3X-FLAG proteins, Gl Ter, and
GPx1 Ter mRNA localize to P-bodies as shown by the co-
localization with CFP-hDCP1a. This result is similar to what
we observed when cells were treated with NMDI 1. Interest-
ingly, YFP-hSMG5 was not detected in P-bodies, suggesting
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1153
Figure 5. PTC-containing mRNAs accumulate in P-Bodies under conditions of NMD inhibition. (Α) HeLa cells were transfected with pmCMV-Gl Ter and an expression vector encoding a P-body component: pCI-neo-FLAG-hUpf1, pYFP-hSmg6, pYFP-hSmg7, pcDNA3-hUpf3-FLAG, pGFP-Ge1, or pGFP-hCcr4. 24 h after transfection, DMSO(−) or 5 μM NMDI 1 was added to the culture medium for 20 h. (B) U2OS cells were transfected with a pGl-Ter MS2 plasmid and a pCFP-hDcp1a expression vector. These cells were used for their tendency to highly express transfected genes, which is crucial for this approach. Cells were treated as described in A. Nuclei are shown in the merge by Hoechst staining in blue. The white squares are magnifi cations of cell areas. Bars, 10 μm.
JCB • VOLUME 178 • NUMBER 7 • 2007 1154
that by destabilizing the interaction between hSMG5 and hUPF1,
hSMG5 is excluded from P-bodies. Additionally, expression of
either HA-SMG5dCT or HA-SMG5DA induced accumulation
of FLAG-hUPF1, hUPF3-FLAG, hUPF3X-FLAG proteins,
Gl Ter, and GPx1 Ter mRNA into P-bodies (Fig. 8, B and C,
respectively). Thus, by specifi cally blocking the dephosphor-
ylation of hUPF1, NMD factors and substrates concentrate
into P-bodies.
Figure 6. Wild-type mRNAs do not accumulate in P-bodies when NMD is inhibited by NMDI 1. (A) HeLa cells were transfected with pmCMV-Gl Norm and an expression vector encoding a P-body component: pYFP-hSmg6 or pGFP-Ge1. Cells were submitted to the same treatment as in Fig. 5 A. (B) Same as in Fig. 5 B except that cells were transfected with a pGl Norm MS2 plas-mid rather than pGl-Ter MS2. The white squares are magnifi cations of cell areas. Bars, 10 μm.
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1155
DiscussionIn this study, we have characterized an indole derivative, NMDI 1,
as an NMD inhibitor. This molecule has allowed us to study
specifi c steps of NMD. The power of this approach lies in the
ability to freeze NMD at a step when hUPF1 and hUPF3/3X are
detected in P-bodies (Figs. 3 and 4). Using biochemical and cel-
lular biology approaches, we have determined the precise event
blocked by NMDI 1 and established that this inhibitor prevents
the interactions between hUPF1 and hSMG5, resulting in the
subsequent exclusion of hSMG5 from P-bodies and the stabili-
zation of hyperphosphorylated isoforms of hUPF1. Unlike other
approaches, such as transfection-mediated down-regulation of
NMD factors, where only a fraction of cells are subject to the
inhibition of NMD, small chemical molecules have the ability
to diffuse across the cell membrane and affect most cells in
culture. Therefore, such inhibitors should enable NMD to be
inhibited in more physiologically complex environments such
as tissue or multicellular organisms to study NMD mechanism
in vivo and to evaluate their potential therapeutic capacities
(Kuzmiak and Maquat, 2006).
As in yeast (Sheth and Parker, 2006), PTC-containing
mRNAs, hUPF1, and hUPF3/3X proteins are found in P-bodies
of mammalian cells when NMD is prevented (Figs. 4, 5, and 8).
Figure 7. Down-regulation of hUPF2 or hXRN1 does not lead to the accumulation of NMD factors in P-bodies. (A and C) The effi ciency of the hUPF2 (A) or hXRN1 (C) down-regulation was evaluated by Western blot from 10 μg of total protein. The three leftmost lanes represent 2× dilutions of total extract from HeLa cells that were not transfected. (B) HeLa cells were transfected with pCFP-hDcp1a and either pCI-neo-FLAG-hUPF1, pcDNA3-hUPF3-FLAG, pcDNA3-hUPF3X-FLAG, or pmCMV-Gl Ter in the presence of siRNA luciferase (Luc) or siRNA hUPF2 (Upf2). (D) Same as B except siRNA hXRN1 replaced siRNA hUPF2, and pcDNA3-hUPF3X-FLAG was not tested. The white squares are magnifi cations of cell areas. Bars, 10 μm.
JCB • VOLUME 178 • NUMBER 7 • 2007 1156
Figure 8. NMDI 1 effect on the localization of NMD factors and substrates can be reproduced by using hUPF1 or hSMG5 mutants that prevent the phosphorylation of hUPF1. HeLa cells were transfected with pCFP-hDcp1a and pHA-hUpf1dNT (A), pHA-hSmg5dCT (B) or pHA-hSmg5DA (C), and either pCI-neo-FLAG-hUpf1, pcDNA3-hUpf3-FLAG, pcDNA3-hUpf3X-FLAG, pmCMV-Gl-Ter, pmCMV-GPx1-Ter, pYFP-hSmg5 (A), or pCI-neo-FLAG-hUpf1 (B and C). The white squares are magnifi cations of cell areas. Bars, 10 μm.
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1157
Undoubtedly, yeast and human P-bodies share some similarities
in their protein compositions and functions but clear differences
can also be seen. Unlike in yeast, in mammalian cells PTC-
containing mRNAs accumulate in P-bodies, or more precisely
at the periphery of P-bodies, suggesting that P-bodies can be
formed by several compartments. This observation is consistent
with a recent paper (Pillai et al., 2005) showing that RNA can
be localized at the periphery of P-bodies where it might be
stored before being degraded or released from the P-bodies.
Another difference between yeast and mammalian P-bodies is
that a down-regulation of hUPF2 in mammals does not lead to
the accumulation of hUPF1 into P-bodies (Fig. 7) as it does in
yeast (Sheth and Parker, 2006). Our results suggest that hUPF2
is involved in the NMD process before the transit of NMD fac-
tors and substrates through P-bodies. These differences likely
refl ect a divergence in the process of NMD in yeast and in mam-
malian cells. Surprisingly, when we blocked NMD at a late step
when RNAs are going to be degraded, (i.e., by down-regulating
hXRN1) we did not detect NMD factors in P-bodies but we did
observe PTC-containing mRNAs in P-bodies (Fig. 7). This con-
fi rms that NMD involves traffi cking to P-bodies and suggests
that by inhibiting the RNA degradation step, we did not prevent
the recycling of NMD factors from P-bodies.
Interestingly, we did not observe any differences in the
cellular distribution and in the protein composition between
P-bodies containing Gl Ter mRNA and those containing GPx1
Ter mRNA, even though these two mRNAs are subject to
nucleus-associated or cytoplasmic NMD, respectively. However,
we cannot exclude the possibility that NMDI 1 would freeze a
series of dynamic events that occurs during NMD and that this
would induce a drift of nucleus-associated P-bodies to the cyto-
plasm. It is also possible that P-bodies with nucleus-associated
NMD substrates and P-bodies with cytoplasmic NMD sub-
strates have different biochemical or physical properties that
would lead to the cosedimentation of one with the nuclear frac-
tion and of the other with the cytoplasmic fraction. Further
investigations will be necessary to clarify this point.
NMDI 1 allowed us to show that P-bodies display a
large degree of variability in their NMD factor composition.
For example, whereas hSMG6 colocalizes with CFP-hDCP1a
or FLAG-hUPF1 when NMD is inhibited (Fig. 3 and not de-
picted), it only shows a partial overlap with CFP-hDCP1a in
DMSO-treated cells (Fig. 3) and no colocalization with LSM4
(Unterholzner and Izaurralde, 2004). As another example, we
detected hUPF3/3X-FLAG proteins in P-bodies positive for
YFP-hSMG6 or YFP-hSMG7 but not in P-bodies containing
GFP-GE1 (Fig. 4). Similarly, PTC-containing mRNA is found
in all P-bodies holding hUPF3/3X-FLAG proteins and in most
P-bodies containing FLAG-hUPF1, YFP-hSMG6, or YFP-hSMG7,
but rarely or never in P-bodies stained with GFP-hCCR4 or
GFP-GE1 (Figs. 5 and S4). Because the factors that do not
colocalize upon NMDI 1 treatment, such as hUPF3/3X (or PTC-
containing mRNAs) and GE1, still form foci that colocalize with
other P-body markers (such as hSMG7), these data suggest that
P-bodies can exist in several “fl avors” or forms that would dif-
fer in protein composition (at least in mammalian cells). These
variations in P-body composition could refl ect different func-
tional states. In this view, hUPF3/3X, hSMG6, hSMG7, hUPF1,
and hDCP1a would be involved at early steps of RNA process-
ing in P-bodies or, as has been recently proposed, would nucle-
ate the formation of P-bodies on the PTC-containing mRNA
(Franks and Lykke-Andersen, 2007). hCCR4 would then join
the structure, followed by GE1 that would induce degradation
of the PTC-containing mRNA and recycling of hUPF3/3X.
Then other NMD factors are recycled for a new turn of NMD
(Fig. 9). This evolution in P-body composition could arise by
fusion of different subcategories of P-bodies, by the shuttling of
individual components, or by a combination of these processes.
An attractive approach to answer these questions would be to
characterize NMD inhibitors that target other steps than the de-
phosphorylation of hUPF1.
Materials and methodsAll results presented in this article are representative from at least three independent experiments.
Chemical molecules libraryAll the polycyclic indole compounds studied in this paper issue from the Institut Curie–Centre National de la Recherche Scientifi que compound library. Each compound was suspended at 20 mg/ml in DMSO and then prepared at a 5-μM working dilution in 10% DMSO (vol/vol). The synthe-sis and the purifi cation of these compounds has been described previously (Rivalle et al., 1981).
ConstructsGl Norm and Ter fused to MS2 binding sites constructs (Figs. 5 B and 6 B) were obtained by PCR amplifi cation from Gl wild-type and NS39 constructs (provided by M. Hentze and N. Gehring, European Molecular Biology Laboratory, Heidelberg, Germany; Thermann et al., 1998) using a sense
Figure 9. A model of P-body composition dynamics. (1) At the early stage, phosphorylated hUPF1, hUPF3/3X, hSMG5, hSMG6, hSMG7, and PTC-containing mRNAs transit to P-bodies. (2) hCCR4 accumulates later, followed by GE1, which induces the degradation of PTC-containing mRNAs, and (3) the recycling of hUPF3/3X. Finally, other NMD factors will be recycled, in particular hUPF1.
JCB • VOLUME 178 • NUMBER 7 • 2007 1158
primer (5′-G C A A C C T C A A G C T T A C A C C A T G G T G C A C C T G A C -3′) and an antisense primer (5′-A G A A A G C A G A T C T G C T T A G T G A T A C T T G T G -3′). The amplifi ed fragment was cloned in HindIII–BglII of a modifi ed pRSVbgal plasmid containing 24× MS2 sites (Fusco et al., 2003).
NMD measurements by RT-PCRHeLa cells were cultured in 60-mm dishes in DME (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum at 37°C and 5% CO2. 106 cells were cotransfected with 3 μg of test plasmid pmCMV-Gl (Norm or 39Ter; Sun et al., 1998) or 3 μg of test plasmid pmCMV-GPx1 (Norm or 46Ter; Moriarty et al., 1998) and 1 μg of reference plasmid phCMV-MUP (all provided by L. Maquat, University of Rochester, Rochester, NY; Belgrader and Maquat, 1994) using Lipofectamine Plus reagent (Invitrogen) accord-ing to the manufacturer’s instructions. 24 h after transfection, cells were treated for 20 h with 5 μM of chemical molecules or 0.01% DMSO (vol/vol) as a control. Total RNA was purifi ed using the TRI reagent (Sigma-Aldrich), and Gl, GPx1, and MUP mRNA were reverse transcribed before amplifi cation by PCR in the presence of 32P-radiolabeled dCTP. The PCR condi-tions and analysis method have been previously described (Ishigaki et al., 2001). PCR products were quantifi ed with an imaging system (Typhoon 9200; GE Healthcare).
Luciferase activity for translation effi ciency assayHeLa cells were transfected with 2 μg pFluc and 1 μg pRluc. 24 h after transfection, cells were incubated with 0.01% DMSO (vol/vol), 5 μM NMDI 1 for 20 h, or 100 μg/ml CHX for 4 h before being harvested. Luciferase activity was quantifi ed on 2 × 105 cell equivalent on a lumino-meter (MicroLumat LB 96P; Berthold Technologies) using the Dual-Glo lucif-erase assay kit (Promega) according to the manufacturer’s instructions. The luciferase activity was normalized according to the level of Fluc and Rluc mRNA level measured by RPA.
Luciferase activity for miRNA decay pathway integrityCells were grown in six-well plates and transfected with Lipofectamine Plus reagent using 50 ng RLperfect RNA reporter (Pillai et al., 2005), 200 ng pGl3 plasmid coding for Fluc, and 4 μg pTzU6 plasmid (Good et al., 1997). 24 h after transfection, NMDI 1 was added to corresponding wells. 48 h after transfection, luciferase activity was measured with a Dual-Glo luciferase assay kit according to the manufacturer‘s instructions.
Measure of FLAG-hUPF1 phosphorylation level by 2D gel analysis106 293T cells were transfected with 1 μg pCI-neo-FLAG-hUpf1 (provided by L. Maquat; Sun et al., 1998) using Lipofectamine Plus reagent accord-ing to the manufacturer’s instructions. After 12 h, serum was removed from the culture medium for 24 h before adding 5 μM NMDI 1 or 0.01% DMSO (vol/vol) as a control for 3 h at 37°C and 5% CO2. 10% serum was added back for 1 h at 37°C and 5% CO2. Total proteins were purifi ed in a lysis buffer containing 8 M urea, 2% CHAPS, and 40 mM Tris base. The fi rst dimension was performed according to the protocol guide from GE Healthcare for 2D electrophoresis with immobilized pH gradient. 18-cm Immobiline DryStrip, pH 3–10 (GE Healthcare), was used to separate pro-teins according to their isoelectric point. The second dimension was done by loading the fi rst dimension on a 10% SDS-PAGE. Finally, proteins were transferred to a nitrocellulose membrane before incubation with 1 μg of mouse anti–α-FLAG antibody (Sigma-Aldrich) in TBS containing 0.05% Tween overnight at 4°C followed by incubation with a peroxidase-conjugated goat anti–mouse antibody (Pierce Chemical Co.). Proteins were then detected using SuperSignal West Femto maximum sensitivity substrate (Pierce Chemical Co.).
Tethering assayThis experiment was performed as described previously (Hosoda et al., 2005).
Immunofl uorescence, FISH assays, and image analysisHeLa cells were cultured on 12-mm glass coverslips in 10% DME (vol/vol) FBS. Cells (105) were transiently transfected with 500 ng of plasmids pGFP-Ge1 (provided by D. Bloch, Harvard Medical School, Charleston, MA; Yu et al., 2005), pYFP-hSmg5, pYFP-hSmg6, pYFP-hSmg7 (all provided by E. Izaurralde, Max Planck Institute for Developmental Biology, Tuebingen, Germany; Unterholzner and Izaurralde, 2004), pGFP-hCcr4, pCFP-hDcp1a (both provided by B. Séraphin, Centre National de la Recherche Scienti-fi que, Montpellier, France; Cougot et al., 2004), pCI-neo-FLAG-hUpf1 (Sun et al., 1998), pcDNA3-hUpf3-FLAG, pcDNA3-hUpf3X-FLAG (both provided by L. Maquat; Lykke-Andersen et al., 2000), pmCMV-Gl (Norm or 39Ter; Sun et al., 1998), pmCMV-GPx1 (Norm or 46Ter; Moriarty et al., 1998), pHA-hUpf1dNT, pHA-hSmg5dCT, or pHA-hSmg5DA (all three provided by
S. Ohno and A. Yamashita, Yokohama City University School of Medicine, Yokohama, Japan; Ohnishi et al., 2003). 24 h after transfection, cells were treated with 5 μM NMDI 1 or 0.01% DMSO (vol/vol) as a control. 20 h later cells were fi xed using formalin solution (Sigma-Aldrich) for 10 min at room temperature and permeabilized in 70% ethanol overnight at 4°C. For immunofl uorescence assays, fi xed cells were incubated with a mouse anti-FLAG antibody (Sigma-Aldrich) for 2 h at room temperature, washed three times with PBS, and incubated with Cy3- or FITC-conjugated goat anti–mouse antibody (Jackson ImmunoResearch Laboratories) for 1 h at room temperature. Finally, cells were washed three times with PBS and incubated with 2 ng/μl of Hoechst stain (Sigma-Aldrich) for 2 min at room temperature. For FISH experiments, fi xed cells were incubated in a pre-hybridization buffer (125 μg/ml tRNA, 500 μg/ml herring DNA, 1 mg/ml BSA, 0.1 g/ml dextran sulfate, 50% formamide, and 2× sodium saline citrate [SSC] buffer) at 37°C for 1 h in a tissue culture incubator. Fixed cells were incubated overnight in a tissue culture incubator with a hybridiza-tion buffer (prehybridization buffer with Cy3-labeled probes), washed three times in 2× SSC at 37°C and three times in 1× SSC at room temperature, and incubated with 2 ng/μl of Hoechst stain (Sigma-Aldrich) for 2 min at room temperature. A Cy3 5′ and 3′ end–labeled probe (5′-C G A T C T G C G-T T C T A C G G T G G T -3′) was used to detect Gl or GPx1 mRNA Ter. For Figs. 5 B and 6 B, the MS2 probe sequence was 5′-AT*G T C G A C C T G C A G A C A T *-G G G T G A T C C T C A T *G T T T T C T A G G C A A T T *A-3′ (the asterisks indicate an internal Cy3 modifi cation).
Fixed cells were observed in V E C T A S H I E L D mounting medium (Vector Laboratories) with a microscope (DMRA; Leica) with an oil objective (PL APO 63×, NA 1.32; Leica) and A4 (for Hoechst stain), GFP, and Y3 (for Cy3) fi lter sets (Leica). Images were captured at 0.3-μm intervals along the z axis using a piezzo stepper (E662 LVPTZ amplifi er; Servo Products) and a cooled charge-coupled device camera (MicroMAX, 1,300 × 1,030 pix-els, RS; Princeton Instruments Inc.) driven by MetaMorph v6.2 (Universal Imaging Corp.). Pixel sizes were 106 × 106 nm and voxel sizes were 106 × 106 × 100 nm. For deconvolution and image reconstruction, xyz image stacks of fi xed cells were processed using deconvolution software (Huy-gens 2.3; Scientifi c Volume Imaging) using a maximum likelihood estima-tion algorithm. 3D restored stacks were processed with software (Imaris 4; Bitplane) for volume rendering and quantifi cation.
Down-regulation of hUPF2 or hXRN1105 HeLa cells were transfected with 100 nM siRNA hUpf2 (Eurogentec; Kim et al., 2005) using a JetPEI reagent (PolyPlus Transfection) for 48 h be-fore being harvested. The down-regulation effi ciency was then analyzed by Western blot or by immunofl uorescence.
Down-regulation of hXRN1 (provided by J. Lykke-Andersen, Univer-sity of Colorado, Boulder, CO) has been previously described (Cougot et al., 2004).
Immunopurifi cation and Western blot analysishUPF1 immunopurifi cation and Western blot analysis were performed according to the protocol described in Lejeune and Maquat (2004) using rabbit anti–α-hUPF1 antibody (provided by S. Ohno and A. Yamashita; Ohnishi et al., 2003). Western blot analyses were performed using 1:250 rabbit anti–α-hUPF1, α-hSMG5, α-hSMG6, or α-hSMG7 antibodies (pro-vided by S. Ohno and A. Yamashita; Ohnishi et al., 2003); 1:1,000 rab-bit α-hUPF3/3X (Ishigaki et al., 2001); or 1:1,000 mouse anti–α-tubulin (Sigma-Aldrich). For Fig. 2, Western blot analysis was done with 1:1,000 rabbit anti–α-hUPF1, 1:1,000 rabbit anti–α-hUPF2, and 1:1,000 rabbit anti–α-hUPF3/3X (Ishigaki et al., 2001). Proteins were detected using SuperSignal West Pico chemiluminescent substrate or SuperSignal West Femto maximum sensitivity substrate (Pierce Chemical Co.).
Online supplemental materialFig. S1 shows that pre-mRNA splicing and mRNA translation are not af-fected by NMDI 1. Fig. S2 confi rms the results of Fig. 2 by RPA approach and presents the evidence of the accumulation of FLAG-hUPF1 in P-bodies of 293T cells in the presence of NMDI 1 and serum. Fig. S3 shows that endogenous phosphorylated hUpf1, hSmg6, and hSmg7 are detected in P-bodies in the presence of NMDI 1. Fig. S4 demonstrates that GPx1-Ter mRNA is present in P-bodies when NMD is blocked by NMDI 1. Finally, Table S1 shows the structures of compounds that have been used in this study. Online supplemental material is available at http://www.jcb.org/content/full/jcb.200611086/DC1.
The authors sincerely thank Dr. Lynne Maquat for the pmCMV-Gl (Norm and Ter), pmCMV-GPx1 (Norm and Ter), phCMV-MUP, pCI-neo-FLAG-hUpf1, pcDNA3-hUpf3-FLAG, and pcDNA3-hUpf3X-FLAG plasmids and anti-hUPF1, anti-hUPF2,
PROCESSING BODY COMPONENT DYNAMICS • DURAND ET AL. 1159
and anti-hUPF3/3X antibodies; Dr. Elisa Izaurralde for providing the pYFP-hSmg5, pYFP-hSmg6, and pYFP-hSmg7 plasmids; and Dr. Jens Lykke-Andersen for providing plasmids for tethering experiments and the anti-hXRN1 antibody. We want also to thank Dr. Donald Bloch for providing the pGFP-Ge1 plasmid; Dr. Matthias Hentze and Dr. Niels Gehring for the Gl wild-type and NS39 constructs; and Dr. Shigeo Ohno and Dr. Akia Yamashita for providing us with anti-hSMG5, anti-hSMG6, anti-hSMG7, the anti-phosphorylated isoform of hUPF1 and anti-hUPF1 antibodies, and the pHA-hUpf1dNT, pHA-Smg5DA, and pHA-Smg5dCT vectors. We further thank Dr. Bertrand Séraphin for the pCFP-hDcp1a and pGFP-hCcr4 plasmids. Finally, we would like to thank Dr. Johann Soret, Dr. Oliver Mühlemann, and Dr. Naomi Taylor for critical read-ing of the manuscript and the Montpellier RIO Imaging facility for help in microscopy analysis.
S. Durand was supported by a graduate fellowship from the Ministère Délégué à la Recherche et aux Technologies. F. Lejeune was supported for nine months by the Fondation pour la recherche médicale. This work was sup-ported by a grant from the Association Française contre les Myopathies to F. Lejeune and a grant from Agence Nationale pour la Recherche (ANR 05-BLAN-0261-01) to J. Tazi.
Submitted: 15 November 2006Accepted: 22 August 2007
ReferencesBashkirov, V.I., H. Scherthan, J.A. Solinger, J.M. Buerstedde, and W.D. Heyer.
1997. A mouse cytoplasmic exoribonuclease (mXRN1p) with preference for G4 tetraplex substrates. J. Cell Biol. 136:761–773.
Belgrader, P., and L.E. Maquat. 1994. Nonsense but not missense mutations can decrease the abundance of nuclear mRNA for the mouse major urinary protein, while both types of mutations can facilitate exon skipping. Mol. Cell. Biol. 14:6326–6336.
Brengues, M., D. Teixeira, and R. Parker. 2005. Movement of eukaryotic mRNAs between polysomes and cytoplasmic processing bodies. Science. 310:486–489.
Chen, C.Y., and A.B. Shyu. 2003. Rapid deadenylation triggered by a nonsense codon precedes decay of the RNA body in a mammalian cytoplasmic nonsense-mediated decay pathway. Mol. Cell. Biol. 23:4805–4813.
Chiu, S.Y., G. Serin, O. Ohara, and L.E. Maquat. 2003. Characterization of human Smg5/7a: a protein with similarities to Caenorhabditis elegans SMG5 and SMG7 that functions in the dephosphorylation of Upf1. RNA. 9:77–87.
Conti, E., and E. Izaurralde. 2005. Nonsense-mediated mRNA decay: molecular insights and mechanistic variations across species. Curr. Opin. Cell Biol. 17:316–325.
Cougot, N., S. Babajko, and B. Seraphin. 2004. Cytoplasmic foci are sites of mRNA decay in human cells. J. Cell Biol. 165:31–40.
Couttet, P., and T. Grange. 2004. Premature termination codons enhance mRNA decapping in human cells. Nucleic Acids Res. 32:488–494.
Czaplinski, K., M.J. Ruiz-Echevarria, S.V. Paushkin, X. Han, Y. Weng, H.A. Perlick, H.C. Dietz, M.D. Ter-Avanesyan, and S.W. Peltz. 1998. The sur-veillance complex interacts with the translation release factors to enhance termination and degrade aberrant mRNAs. Genes Dev. 12:1665–1677.
Fenger-Gron, M., C. Fillman, B. Norrild, and J. Lykke-Andersen. 2005. Multiple processing body factors and the ARE binding protein TTP activate mRNA decapping. Mol. Cell. 20:905–915.
Franks, T., and J. Lykke-Andersen. 2007. TTP and BRF proteins nucleate pro-cessing body formation to silence mRNAs with AU-rich elements. Genes Dev. 21:719–735.
Fukuhara, N., J. Ebert, L. Unterholzner, D. Lindner, E. Izaurralde, and E. Conti. 2005. SMG7 is a 14-3-3-like adaptor in the nonsense-mediated mRNA decay pathway. Mol. Cell. 17:537–547.
Fusco, D., N. Accornero, B. Lavoie, S.M. Shenoy, J.M. Blanchard, R.H. Singer, and E. Bertrand. 2003. Single mRNA molecules demonstrate probabilis-tic movement in living mammalian cells. Curr. Biol. 13:161–167.
Gehring, N.H., J.B. Kunz, G. Neu-Yilik, S. Breit, M.H. Viegas, M.W. Hentze, and A.E. Kulozik. 2005. Exon-junction complex components specify dis-tinct routes of nonsense-mediated mRNA decay with differential cofactor requirements. Mol. Cell. 20:65–75.
Good, P.D., A.J. Krikos, S.X. Li, E. Bertrand, N.S. Lee, L. Giver, A. Ellington, J.A. Zaia, J.J. Rossi, and D.R. Engelke. 1997. Expression of small, thera-peutic RNAs in human cell nuclei. Gene Ther. 4:45–54.
He, F., X. Li, P. Spatrick, R. Casillo, S. Dong, and A. Jacobson. 2003. Genome-wide analysis of mRNAs regulated by the nonsense-mediated and 5′ to 3′ mRNA decay pathways in yeast. Mol. Cell. 12:1439–1452.
Hosoda, N., Y.K. Kim, F. Lejeune, and L.E. Maquat. 2005. CBP80 promotes interaction of Upf1 with Upf2 during nonsense-mediated mRNA decay in mammalian cells. Nat. Struct. Mol. Biol. 12:893–901.
Ingelfi nger, D., D.J. Arndt-Jovin, R. Luhrmann, and T. Achsel. 2002. The human LSm1-7 proteins colocalize with the mRNA-degrading enzymes Dcp1/2 and Xrnl in distinct cytoplasmic foci. RNA. 8:1489–1501.
Ishigaki, Y., X. Li, G. Serin, and L.E. Maquat. 2001. Evidence for a pioneer round of mRNA translation: mRNAs subject to nonsense-mediated decay in mammalian cells are bound by CBP80 and CBP20. Cell. 106:607–617.
Kashima, I., A. Yamashita, N. Izumi, N. Kataoka, R. Morishita, S. Hoshino, M. Ohno, G. Dreyfuss, and S. Ohno. 2006. Binding of a novel SMG-1-Upf1-eRF1-eRF3 complex (SURF) to the exon junction complex triggers Upf1 phosphorylation and nonsense-mediated mRNA decay. Genes Dev. 20:355–367.
Kedersha, N., G. Stoecklin, M. Ayodele, P. Yacono, J. Lykke-Andersen, M.J. Fitzler, D. Scheuner, R.J. Kaufman, D.E. Golan, and P. Anderson. 2005. Stress granules and processing bodies are dynamically linked sites of mRNP remodeling. J. Cell Biol. 169:871–884.
Kim, Y.K., L. Furic, L. Desgroseillers, and L.E. Maquat. 2005. Mammalian Staufen1 recruits Upf1 to specifi c mRNA 3′UTRs so as to elicit mRNA decay. Cell. 120:195–208.
Kuzmiak, H.A., and L.E. Maquat. 2006. Applying nonsense-mediated mRNA decay research to the clinic: progress and challenges. Trends Mol. Med. 12:306–316.
Lejeune, F., and L.E. Maquat. 2004. Immunopurifi cation and analysis of protein and RNA components of mRNP in mammalian cells. Methods Mol. Biol. 257:115–124.
Lejeune, F., and L.E. Maquat. 2005. Mechanistic links between nonsense- me diated mRNA decay and pre-mRNA splicing in mammalian cells. Curr. Opin. Cell Biol. 17:309–315.
Lejeune, F., X. Li, and L.E. Maquat. 2003. Nonsense-mediated mRNA decay in mammalian cells involves decapping, deadenylating, and exonucleolytic activities. Mol. Cell. 12:675–687.
Lykke-Andersen, J., M.D. Shu, and J.A. Steitz. 2000. Human Upf proteins target an mRNA for nonsense-mediated decay when bound downstream of a termination codon. Cell. 103:1121–1131.
Lykke-Andersen, J., M.D. Shu, and J.A. Steitz. 2001. Communication of the position of exon-exon junctions to the mRNA surveillance machinery by the protein RNPS1. Science. 293:1836–1839.
Maquat, L.E. 2004a. Nonsense-mediated mRNA decay: a comparative analysis of different species. Curr. Genomics. 5:175–190.
Mendell, J.T., C.M. ap Rhys, and H.C. Dietz. 2002. Separable roles for rent1/hUpf1 in altered splicing and decay of nonsense transcripts. Science. 298:419–422.
Mendell, J.T., N.A. Sharifi , J.L. Meyers, F. Martinez-Murillo, and H.C. Dietz. 2004. Nonsense surveillance regulates expression of diverse classes of mamma-lian transcripts and mutes genomic noise. Nat. Genet. 36:1073–1078.
Moriarty, P.M., C.C. Reddy, and L.E. Maquat. 1998. Selenium defi ciency re-duces the abundance of mRNA for Se-dependent glutathione peroxidase 1 by a UGA-dependent mechanism likely to be nonsense codon-mediated decay of cytoplasmic mRNA. Mol. Cell. Biol. 18:2932–2939.
Ohnishi, T., A. Yamashita, I. Kashima, T. Schell, K.R. Anders, A. Grimson, T. Hachiya, M.W. Hentze, P. Anderson, and S. Ohno. 2003. Phosphorylation of hUPF1 induces formation of mRNA surveillance complexes contain-ing hSMG-5 and hSMG-7. Mol. Cell. 12:1187–1200.
Page, M.F., B. Carr, K.R. Anders, A. Grimson, and P. Anderson. 1999. SMG-2 is a phosphorylated protein required for mRNA surveillance in Caenorhabditis elegans and related to Upf1p of yeast. Mol. Cell. Biol. 19:5943–5951.
Pal, M., Y. Ishigaki, E. Nagy, and L.E. Maquat. 2001. Evidence that phosphoryl-ation of human Upfl protein varies with intracellular location and is mediated by a wortmannin-sensitive and rapamycin-sensitive PI 3-kinase-related kinase signaling pathway. RNA. 7:5–15.
Pillai, R.S., S.N. Bhattacharyya, C.G. Artus, T. Zoller, N. Cougot, E. Basyuk, E. Bertrand, and W. Filipowicz. 2005. Inhibition of translational initiation by Let-7 MicroRNA in human cells. Science. 309:1573–1576.
Rehwinkel, J., I. Letunic, J. Raes, P. Bork, and E. Izaurralde. 2005. Nonsense-mediated mRNA decay factors act in concert to regulate common mRNA targets. RNA. 11:1530–1544.
Rivalle, C., C. Ducrocq, J.M. Lhoste, F. Wendling, E. Bisagni, and J.C. Chermann. 1981. 11H-Pyrido [3′,2′:4,5]pyrrolo[2,3-g]isoquinoléines (aza-7 ellipticines) substituées sur leur position 6. Tetrahedron. 37:2097–2103.
Serin, G., A. Gersappe, J.D. Black, R. Aronoff, and L.E. Maquat. 2001. Identifi cation and characterization of human orthologues to Saccharomyces cerevisiae Upf2 protein and Upf3 protein (Caenorhabditis elegans SMG-4). Mol. Cell. Biol. 21:209–223.
JCB • VOLUME 178 • NUMBER 7 • 2007 1160
Sheth, U., and R. Parker. 2003. Decapping and decay of messenger RNA occur in cytoplasmic processing bodies. Science. 300:805–808.
Sheth, U., and R. Parker. 2006. Targeting of aberrant mRNAs to cytoplasmic processing bodies. Cell. 125:1095–1109.
Soret, J., N. Bakkour, S. Maire, S. Durand, L. Zekri, M. Gabut, W. Fic, G. Divita, C. Rivalle, D. Dauzonne, et al. 2005. Selective modifi cation of alterna-tive splicing by indole derivatives that target serine-arginine-rich protein splicing factors. Proc. Natl. Acad. Sci. USA. 102:8764–8769.
Sun, X., H.A. Perlick, H.C. Dietz, and L.E. Maquat. 1998. A mutated human homologue to yeast Upf1 protein has a dominant-negative effect on the decay of nonsense-containing mRNAs in mammalian cells. Proc. Natl. Acad. Sci. USA. 95:10009–10014.
Sureau, A., R. Gattoni, Y. Dooghe, J. Stevenin, and J. Soret. 2001. SC35 autoregu-lates its expression by promoting splicing events that destabilize its mRNAs. EMBO J. 20:1785–1796.
Teixeira, D., U. Sheth, M.A. Valencia-Sanchez, M. Brengues, and R. Parker. 2005. Processing bodies require RNA for assembly and contain nontrans-lating mRNAs. RNA. 11:371–382.
Thermann, R., G. Neu-Yilik, A. Deters, U. Frede, K. Wehr, C. Hagemeier, M.W. Hentze, and A.E. Kulozik. 1998. Binary specifi cation of nonsense codons by splicing and cytoplasmic translation. EMBO J. 17:3484–3494.
Tourriere, H., K. Chebli, L. Zekri, B. Courselaud, J.M. Blanchard, E. Bertrand, and J. Tazi. 2003. The RasGAP-associated endoribonuclease G3BP assembles stress granules. J. Cell Biol. 160:823–831.
Unterholzner, L., and E. Izaurralde. 2004. SMG7 acts as a molecular link be-tween mRNA surveillance and mRNA decay. Mol. Cell. 16:587–596.
van Dijk, E., N. Cougot, S. Meyer, S. Babajko, E. Wahle, and B. Seraphin. 2002. Human Dcp2: a catalytically active mRNA decapping enzyme located in specifi c cytoplasmic structures. EMBO J. 21:6915–6924.
Wollerton, M.C., C. Gooding, E.J. Wagner, M.A. Garcia-Blanco, and C.W. Smith. 2004. Autoregulation of polypyrimidine tract binding protein by alterna-tive splicing leading to nonsense-mediated decay. Mol. Cell. 13:91–100.
Yamashita, A., T. Ohnishi, I. Kashima, Y. Taya, and S. Ohno. 2001. Human SMG-1, a novel phosphatidylinositol 3-kinase-related protein kinase, associates with components of the mRNA surveillance complex and is involved in the regulation of nonsense-mediated mRNA decay. Genes Dev. 15:2215–2228.
Yu, J.H., W.H. Yang, T. Gulick, K.D. Bloch, and D.B. Bloch. 2005. Ge-1 is a cen-tral component of the mammalian cytoplasmic mRNA processing body. RNA. 11:1795–1802.
Zhang, J., X. Sun, Y. Qian, and L.E. Maquat. 1998. Intron function in the non-sense-mediated decay of beta-globin mRNA: indications that pre-mRNA splicing in the nucleus can infl uence mRNA translation in the cytoplasm. RNA. 4:801–815.
